Landscape genetics has undoubtedly generated fundamental insights into the dynamics of rabies but has arguably yet to make major contributions to its control. The preceding discussion illustrated potential avenues for exploration and here we aim to more clearly define a landscape genetics research agenda that could directly benefit the planning or implementation of programmes that aim to control or eliminate rabies.
Considerable progress has been made in the development of blueprints and operational toolkits for rabies control (see “Blueprint for rabies prevention and control,” August 2010, http://www.rabiesblueprint.com/
, and Lembo et al.
). The rabies blueprint presents a major step in the global fight against rabies, providing guidelines and economically feasible strategies to aid policy makers and local communities seeking to embark on rabies intervention and control measures. Landscape genetics research constitutes an important resource within this multidisciplinary approach to help advise and sustain successful control initiatives.
Vaccination is widely deemed the most effective means of rabies control with demonstrated successes in reducing incidence and eliminating disease even in areas with limited resources (Cleaveland et al.
; Schneider et al.
; Lembo et al.
). Molecular genetics has demonstrated that domestic dogs are critical reservoirs for canine rabies, even in parts of Africa with abundant wildlife populations, and therefore indicate that controlling dog rabies through vaccination should eliminate infection from all other species (Lembo et al.
). By changing the susceptibility of populations, mass vaccinations change the landscape in which rabies circulates, and at sufficiently high levels of vaccination coverage, transmission can be interrupted. There are two mechanisms by which vaccination alters the landscape to control rabies. The first is that vaccination itself creates a barrier of susceptible individuals that block the dispersal of the pathogen across the landscape, and the second is via the reinforcement of existing barriers with vaccination. Using landscape genetics to explore the potential impacts of these aspects of landscape control is a logical next step.
Recurrent rabies epidemics occur across large areas where vaccination programmes are patchy and unsustained (Hampson et al.
), suggesting that a proactive long-term vaccination programme coordinated across political boundaries is required for success. Indeed, the effectiveness of such a programme has been proven by the intensive vaccinations coordinated by the Pan American Health Organization (PAHO) in Latin America over the past few decades: dog rabies has been eliminated in a large portion of the southern continent, and reported cases from other countries are highly localised due to restrictions on transmission pathways from vaccination barriers (Schneider et al.
). Crucially, cases of human rabies have dropped in these areas, reinforcing the importance of effective dog rabies control strategies. Strategically placing a vaccine barrier could maintain freedom from rabies resulting from successful interventions in otherwise landlocked areas.
Environmental or anthropogenic barriers to natural transmission of rabies offer the opportunity to strengthen and smarten vaccination initiatives. Vaccination can be viewed as a form of barrier that impedes rabies spread, and therefore many of the techniques used to draw insights on the permeability of barriers could equally be applied to vaccination programmes. Oral Rabies Vaccination
(ORV) campaigns for wildlife, including bait distribution in proximity to a pre-existing natural barrier (the Appalachian Mountains), and utilising existing environmental features to reinforce control campaigns (Wandeler et al.
) have successfully contained wildlife rabies. Past experience has shown vulnerability to breaches associated with the differential permeability of barriers and emphasizes the need for ongoing and targeted surveillance to enable early detection and swift responses to incursions (Russell et al.
). Yet despite this, few studies have quantified the utility of barriers within a landscape, and there are no guidelines available for the design and implementation of effective cordons sanitaires
for dog rabies.
In addition to barrier studies, modern application of spatial data is allowing us to make increasingly accurate measurements of epidemiological parameters that may affect disease dynamics. For example, Bharti et al.
) demonstrate the use of remote sensing to test for human predictors of disease. They used anthropogenic light from satellite imagery as a measure of seasonal fluctuations of human populations. The observed fluctuation in light intensity (as a measure of population density) correlated to measles distribution and spread in cities in Niger, and provided an accurate, near real-time representation of short-term population fluctuations that may drive pathogen transmission. This approach demonstrates the use of relatively simple proxies for quantifying migration patterns in poorly resourced regions, often the same regions carrying the highest disease burden from dog rabies.
Large-scale interventions are expensive, and adaptive management is often required alongside intervention. Refinements in resource management will ultimately rely on a combination of knowledge from genetics, landscape and host ecology as part of a reactive programme. Specific landscape elements affecting dog rabies spread are generally not known a priori
, so exploration of phylogeographic patterns may identify genetically distinct viral or host populations in areas, which could be targeted for vaccination. However, experience from wildlife rabies suggests that caution is warranted when taking such an approach. Firstly, apparent boundaries between areas dominated by different genetic lineages may have emerged during the initial invasion process and thus do not necessarily reflect areas of low permeability for the virus (Real et al.
; Biek and Real, 2010
). Secondly, while the stability of these phylogeographic domains indicates a lack of mixing between them, this simply suggests that immigrating viruses find it difficult to invade areas with an already established focus, but does not signify the absence of viral immigration per se
. Such areas with putatively self-contained endemic foci could therefore experience a high risk of rabies re-emergence following successful eradication, unless vaccination effort remains high. Whether the second consideration equally applies to dog rabies is currently not clear as pertinent empirical studies are lacking. As with many research problems in landscape genetics, these types of question may be very productively studied using simulation tools (Epperson et al.
). Simulations may indicate the most effective location for vaccine corridors/barriers e.g. vaccination on the far side of a natural barrier (Russell et al.
), whereas metapopulation models may elucidate areas with high connectivity that could be the source of persistence in endemic areas. Quantifying the degree of connectivity between sub-populations is of critical importance to understanding how rabies is maintained across landscapes, and is an integral part of designing effective control interventions and cordons sanitaires
that limit and contain the virus.
Ongoing surveillance is crucial to the long-term success of rabies control, and considerable value can be added to surveillance initiatives through the incorporation of landscape genetics, with the potential to determine the source of incursions and reveal transmission pathways. Epidemiological surveillance may facilitate active case detection and identification of circulating strains, helping to identify areas missed by vaccination. Using a molecular genetics approach during the 2007 FMD outbreak in the UK, allowed swift and effective containment of the outbreak by directing interventions to the critical areas (Cottam et al.
). For countries with limited resources, surveillance in areas with ongoing control programmes and retrospective analysis may be useful for identifying remaining foci of infection to be targeted by vaccination, sources of incursions, or spillover events into or from wildlife. In addition, metapopulation models may be utilized to predict the direction of spread of rabies in naïve populations (Beyer et al.
), directing the location of sentinel points and control measures to ‘hotspots’ of infection (Haydon et al.
) and putative transmission networks based on sampled cases should indicate how well current levels of surveillance are capturing rabies incidence based on inferred missing links between sampled cases.
Rabies containment and control will require ongoing surveillance and sustained control efforts. In time, when rabies incidence is reduced to low levels (the period when control measures often lapse), genetics can provide a means of directing resources to where they are most needed, maintaining a cost-effective approach. Given the small genome of RABV (12Kb), and advances in NGS sequencing, future exploration of phylogeographic patterns utilizing whole genome sequencing is a realistic prospect. This promises to provide appropriately fine genetic resolution for samples collected on small spatio-temporal scales that might otherwise be uninformative. Application of novel techniques in landscape genetics to other RNA viruses such as FMD and Influenza highlights interesting possibilities for uncovering disease dynamics and aiding control directives (Cottam et al
, Bedford et al
). The technology and analytical power are available but have yet to be fully exploited for dog rabies, and thus offer exciting prospects on what can be achieved with their implementation in the future.